Multicast OS Provisioning Reducing Network Bandwidth

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing provisioning methods, such as unicast, face inefficiencies and network congestion when distributing large software applications like operating systems across numerous nodes in a networked computing environment, leading to high bandwidth usage and prolonged provisioning times.

Innovation Solution

A multicast provisioning method where a single stream of packets is broadcasted to a multicast group address, allowing nodes to cache and buffer data, and request missing chunks from peers or the source server, significantly reducing network load and bandwidth usage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If unicast provisioning is used to transmit OS images to each node individually, then reliable delivery can be achieved, but network bandwidth consumption increases significantly and provisioning time extends

Engineering Contradiction:
Improveprovisioning delivery reliabilityVSAvoidnetwork bandwidth consumption
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The patent merges multiple unicast transmission streams into a single multicast stream. Instead of sending separate OS image copies to each node individually, the system transmits one consolidated stream that all nodes can receive simultaneously, thereby reducing network bandwidth consumption while maintaining reliable delivery through acknowledgment mechanisms.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The multicast provisioning system creates a universal transmission channel that serves multiple nodes simultaneously. A single OS image stream can be received by any number of nodes in the multicast group, making the transmission mechanism universally applicable to all nodes rather than requiring individualized unicast connections for each node.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Reliability

If unicast provisioning is used to ensure direct communication with each node, then individualized delivery is achieved, but network traffic congestion occurs when provisioning numerous nodes

Engineering Contradiction:
Improveindividualized delivery reliabilityVSAvoidprovisioning throughput
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

Multiple individualized delivery streams are merged into a single multicast stream that carries the OS image to all nodes simultaneously. This combining approach eliminates network traffic congestion caused by parallel unicast transmissions while maintaining the ability to deliver to each node individually through the shared multicast channel.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

Instead of creating multiple copies of the transmission stream (one per node), the system creates a single copy that is replicated by network infrastructure at multicast routers. Nodes receive copies of the data from the single stream without requiring the source to generate multiple separate streams, thereby improving provisioning throughput.

Inventive Principle:
Principle #26Copying

3Reliability

If the image server transmits a complete OS image to each node, then full provisioning is achieved, but the image server remains occupied until all nodes are provisioned

Engineering Contradiction:
Improvecomplete provisioningVSAvoidimage server occupancy time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The OS image transmission is segmented into multiple phases or chunks that can be transmitted independently. Nodes can receive and process different segments simultaneously, allowing the image server to release occupancy after transmitting each segment rather than remaining occupied until complete provisioning of all nodes is finished.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The image server continuously transmits OS image data to multiple nodes simultaneously through the multicast stream without interruption. This continuous transmission allows the server to serve multiple nodes in parallel, reducing total occupancy time compared to sequential unicast provisioning while ensuring complete provisioning of each node.

Inventive Principle:
Principle #20Continuity of useful action

4Ease of operation

If large OS files are transmitted across networks using unicast, then direct node-to-node communication is established, but network bandwidth is consumed quickly and provisioning time increases

Engineering Contradiction:
Improvedirect communication capabilityVSAvoidnetwork bandwidth usage
Core Design Contradiction:
Ease of operationVSLoss of energy

Solution Approach 1:

Multiple direct communication channels are merged into a single multicast channel that maintains direct communication capabilities with each node while consolidating bandwidth usage. The multicast protocol enables direct delivery to each node address in the group without requiring separate physical or logical channels, thereby reducing overall network bandwidth consumption.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS7764683B2Reliable multicast operating system (OS) provisioning
Publication Date: 2010.07.27 ORACLE AMERICAN INC
  • US7764683B2 patent drawing
  • US7764683B2 patent drawing
  • US7764683B2 patent drawing

AI summary

A method and corresponding system for multicast provisioning of operating systems (OS) at reduced bandwidth. The method includes providing a computing device on a digital communications network. An OS image is stored in the computing device, with the image including data sets or chunks making up one or more phases of a system image. The method continues with associating a single multicast address with each of the data sets. The computing device broadcasts each of the data sets of the image along with the associated multicast address on the network as a repeating loop. The method includes sending a signal from the computing device to a set of nodes on the network to initiate a process for loading the stored image, and in response, operating the nodes to load a multicast installer that listens for the data sets and buffers received data packets in an image cache.